PPI.bio is still in its early development, and so you can expect a few features to be missing and for a few bugs to crop up. Thanks so much for your patience.

You can email joseph.szymborskimail.mcgill.ca with any bugs you come across.

If you want to get updates on new features, consider joining the newsletter.

Thanks for citing!

Joseph Szymborski , Amin Emad, RAPPPID: towards generalizable protein interaction prediction with AWD-LSTM twin networks, Bioinformatics, Volume 38, Issue 16, August 2022, Pages 3958–3967, https://doi.org/10.1093/bioinformatics/btac429

Download the citation:

Proteins and their sequences are obtained from the UniProt Swiss-Prot dataset. The PPI.bio database was last updated July, 2023.

Currently, the only proteome supported is H. sapiens, but more will be added very soon.

There are two phases to this task which will take varying times.

Queue Stage — The first stage is the queue stage. The duration of this stage depends on how many concurrent users are requesting the inference service. This stage can take as little as a second, but can be much, much longer depending on the number of queued inferences jobs.

Inference Stage — Once you are the next in queue, inference can begin. This often takes about six minutes.

If a task is taking too long, you can bookmark the URL and revisit at a later time. Closing the browser tab will not stop the progress of your request.